Mi-Ra Oh

Bayesian Estimation of the Two-Parameter Gamma Distribution

ABSTRACT A Bayesian estimation of the two-parameter gamma distribution is considered under the non informative prior. The Bayesian estimator is obtained by Gibbs sampling. The generation of the shape parameter in the Gibbs sampler is implemented using the adaptive rejection sampling method of Gilks and Wild (1992). Finally, the results of our numerical studies show that the Bayesian estimator using Gibbs sampling along with adaptive rejection sampling outperforms the maximum likelihood and moment based estimators, as well as the other Bayesian estimator.

A prosodic analysis of nonderived-environment blocking

Korean t-palatalization, which creates phonemic affricates, does not apply in nonderived forms, while s, n, 1-palatalization, which creates allophones, does apply in nonderived forms. Previous analyses of the failure of t-palatalization in Korean to apply in nonderived environments, (nonderived-environment blocking (NDEB), to use Kiparskys (1993) term) have tried to account for NDEB effects either in terms of constraints on rule application (e.g., the Revised Alternation Condition and the Strict Cycle Condition) or in terms of contextual underspecification. I will argue that NDEB in fact comes from the prosodically determined environment for t-palatalization: unsyllabified /t, th/are the only consonants to undergo t-palatalization, subject to independently motivated phonological principles (i.e., Extraprosodicity). By doing so, the analysis unifies phonological and morphological ‘derivedness’ and also allows previously unrelated phenomena in Korean to be accounted for in a unified manner, e.g., Consonant Cluster Simplification, Coda Neutralization, S-neutralization, and palatalization.

Missing Values Estimation for Time Course Gene Expression Data Using the Sequential Partial Least Squares Regression Fitting

The size of microarray gene expression data is very big and its observation process is also very complex. Thus missing values are frequently occurred. In this paper we propose the sequential partial least squares(SPLS) regression fitting method to estimate missing values for time course gene expression data that has correlations among observations over time points. The SPLS method is to combine the sequential technique with the partial least squares(PLS) regression fitting method. The usefulness of method proposed is evaluated through some simulation study for three yeast time course data.

Comparison of Functions for Filtering Time Course Gene Expression Data with Flat Patterns

Filtering genes that do not appear to contribute to regulation prior to the statistical analysis of time course gene expression data can reduce the dimensions of data and the possibility of misinterpretation due to noise or lack of variation. In this paper, we compare six different functions for filtering genes with flat pattern under the percentile criterion on an observed sample and that on a bootstrap sample. The result of applying to the yeast cell cycle data shows that the variance function is most similar in both samples.

Bayesian Estimation of the Two-Parameter Kappa Distribution

In this paper a Bayesian estimation of the two-parameter kappa distribution was discussed under the noninformative prior. The Bayesian estimators are obtained by the Gibbs sampling. The generation of the shape parameter and scale parameter in the Gibbs sampler is implemented using the adaptive rejection Metropolis sampling algorithm of Gilks et al. (1995). A Monte Carlo study showed that the Bayesian estimators proposed outperform other estimators in the sense of mean squared error.

Bayesian Parameter Estimation of the Four-Parameter Gamma Distribution

A Bayesian estimation of the four-parameter gamma distribution is considered under the noninformative prior. The Bayesian estimators are obtained by the Gibbs sampling. The generation of the shape/power parameter and the power parameter in the Gibbs sampler is implemented using the adaptive rejection sampling algorithm of Gilks and Wild (1992). Also, the location parameter is generated using the adaptive rejection Metropolis sampling algorithm of Gilks, Best and Tan (1995). Finally, the simulation result is presented.

Bayesian Estimation of the Nakagami-m Fading Parameter

A Bayesian estimation of the Nakagami-m fading parameter is developed. Bayesian estimation is performed by Gibbs sampling, including adaptive rejection sampling. A Monte Carlo study shows that the Bayesian estimators proposed outperform any other estimators reported elsewhere in the sense of bias, variance, and root mean squared error.

Classification of Microarray Gene Expression Data by MultiBlock Dimension Reduction

In this paper, we applied the multiblock dimension reduction methods to the classification of tumor based on microarray gene expressions data. This procedure involves clustering selected genes, multiblock dimension reduction and classification using linear discrimination analysis and quadratic discrimination analysis.

Bayesian Parameter Estimation using the MCMC method for the Mean Change Model of Multivariate Normal Random Variates

In this thesis, Bayesian parameter estimation procedure is discussed for the mean change model of multivariate normal random variates under the assumption of noninformative priors for all the parameters. Parameters are estimated by Gibbs sampling method. In Gibbs sampler, the change point parameter is generated by Metropolis-Hastings algorithm. We apply our methodology to numerical data to examine it.